The rapid development of remote sensing technologies, geophysical prospection, and artificial intelligence has significantly expanded the capacity of archaeologists to investigate landscapes and subsurface archaeological remains using non-invasive methods. Satellite imagery, UAV surveys, and geophysical techniques such as magnetometry and Ground Penetrating Radar (GPR) generate large volumes of heterogeneous spatial data that offer unprecedented opportunities for archaeological discovery and monitoring. At the same time, the increasing complexity and volume of these datasets have created a growing gap between data acquisition and the ability to interpret them effectively. This deliverable addresses this challenge by presenting two complementary technological developments designed to support the analysis and interpretation of complex archaeological datasets: S.A.D.A. (Smart Anomaly Detection Assistant) and AIRS-GEO (AI-Driven Interpretation of Remote Sensing and Heterogeneous Geophysical Data). Together, these tools provide a methodological framework that integrates remote sensing analysis, geophysical data interpretation, and artificial intelligence techniques to enhance archaeological investigation and cultural heritage monitoring. S.A.D.A. is an open-source software platform designed to facilitate the analysis of satellite and UAV imagery for archaeological purposes. The platform integrates multiple analytical techniques—such as spectral anomaly detection, unsupervised clustering, spatial autocorrelation analysis, and dimensionality reduction—within a guided and user-friendly workflow. By organizing these tools into a structured analytical process supported by a graphical user interface, S.A.D.A. reduces the technical barriers often associated with remote sensing analysis and allows archaeologists to explore large datasets more efficiently. The platform functions as an analytical assistant, helping users identify spatial anomalies that may correspond to archaeological features while maintaining transparency and reproducibility in the analytical process. Complementing this approach, AIRS-GEO provides an artificial intelligence–based framework for the interpretation of heterogeneous geophysical datasets. The method combines Self-Organizing Maps (SOM)—an unsupervised neural network technique used for dimensionality reduction and pattern recognition—with Local Indicators of Spatial Autocorrelation (LISA), a spatial statistical method used to identify significant spatial clusters. This integrated workflow enables the extraction of coherent spatial patterns from multidimensional datasets derived from geophysical surveys such as magnetometry and Ground Penetrating Radar. The framework enhances the readability of complex datasets and supports the identification of buried archaeological structures that may be difficult to detect through traditional visual inspection. The complementarity between S.A.D.A. and AIRS-GEO reflects a multiscale approach to archaeological investigation. S.A.D.A. primarily supports the large-scale exploration of archaeological landscapes through remote sensing analysis, enabling the identification of potential areas of archaeological interest. AIRS-GEO, in contrast, focuses on the detailed interpretation of geophysical survey data, providing insights into the spatial organization of subsurface archaeological structures. Together, the two tools contribute to a more integrated analytical workflow that bridges the gap between large-scale landscape analysis and high-resolution subsurface investigation. The applications of these tools extend beyond archaeological research to include preventive archaeology, heritage monitoring, and landscape management. By enabling more efficient screening of large areas and improving the interpretation of geophysical datasets, the proposed methods can support evidence-based decision-making in heritage protection and territorial planning. In conclusion, the developments presented in this deliverable demonstrate how the integration of artificial intelligence, remote sensing, and geophysical prospection can enhance the interpretation of complex archaeological datasets. By providing accessible, transparent, and reproducible analytical tools, S.A.D.A. and AIRS-GEO contribute to advancing non-invasive archaeological methodologies and to strengthening the role of digital technologies in cultural heritage research and management.
Masini et al. (Sat,) studied this question.